MoNuSeg+: Dataset

       This is a multi-organ cell dataset containing cell-level detection bounding boxes and instance-level segmentation masks. Our dataset comes from MICCAI2018 MoNuSeg (Multi-organ Nucleus Segmentation) Challenge.(https://monuseg.grand-challenge.org/).
        As shown in Table 1, the training set consists of 30 images from different hospitals and patients. The size of these images is 1000×1000. These images cover seven organs (breast, liver, kidney, prostate, bladder, colon, and stomach) and different disease states (benign and tumors at different stages), and 21706 annotated nuclear boundaries are present. The testing set includes 14 images and 6697 nuclear. Notably, the testing set contains patches from organs that are not in the training set, which increases the difficulty in the task and challenges the scalability of the deep model.

Table 1. MoNuSeg+ training and testing dataset.

Figure 1. Images and annotations in this dataset.

REQUEST THE DATASET